Face Recognition under Mask Occlusion via Multi-Scale PCAnet and Progressive GAN-based Face Restoration
Abstract
With the widespread application of face recognition technology in public safety, identity authentication, and other fields, face recognition under mask cover has gradually become a research hotspot. Traditional methods rely on manual feature extraction and template matching, making it difficult to deal with the loss of face features caused by occlusion. To accurately recognize faces under mask occlusion, an improved principal component analysis network based on multi-scale feature fusion (Multi-Scale PCANet, MPCANet) and a progressively trained multi-scale face restoration generative adversarial network (Multi-Scale Face Restoration Generative Adversarial Network, MFR-GAN) are proposed for face biometric recognition under masks. MPCANet takes multi-layer convolution to extract features, concatenates different levels of feature information through channels, and divides the feature map into multi-scale blocks through pyramid pooling layers to form multi-scale feature vectors for matching and recognizing occluded faces. MFR-GAN utilizes a progressive training strategy to gradually increase the resolution from low resolution images for training, and introduces discrete wavelet transform to extract high-frequency information and calculate high-frequency loss, optimizing the quality of image generation. The experiment was conducted on the Labeled Faces in the Wild (LFW) dataset, with an average accuracy of 94.24% and a computation time of 0.87 seconds for MPCANet. The MFR-GAN had a realism of 0.95 and a peak signal-to-noise ratio of 30.25dB at a 30% occlusion ratio. In practical applications, the new model achieved a recognition accuracy of 98.68%, a false recognition rate of 1.32%, and a recognition rate of 97.58% under occlusion conditions, demonstrating high efficiency and accuracy. This method provides an efficient solution for biometric recognition in mask occlusion scenarios, which is helpful for identity authentication.
DOI:
https://doi.org/10.31449/inf.v49i32.9216Downloads
Published
How to Cite
Issue
Section
License
I assign to Informatica, An International Journal of Computing and Informatics ("Journal") the copyright in the manuscript identified above and any additional material (figures, tables, illustrations, software or other information intended for publication) submitted as part of or as a supplement to the manuscript ("Paper") in all forms and media throughout the world, in all languages, for the full term of copyright, effective when and if the article is accepted for publication. This transfer includes the right to reproduce and/or to distribute the Paper to other journals or digital libraries in electronic and online forms and systems.
I understand that I retain the rights to use the pre-prints, off-prints, accepted manuscript and published journal Paper for personal use, scholarly purposes and internal institutional use.
In certain cases, I can ask for retaining the publishing rights of the Paper. The Journal can permit or deny the request for publishing rights, to which I fully agree.
I declare that the submitted Paper is original, has been written by the stated authors and has not been published elsewhere nor is currently being considered for publication by any other journal and will not be submitted for such review while under review by this Journal. The Paper contains no material that violates proprietary rights of any other person or entity. I have obtained written permission from copyright owners for any excerpts from copyrighted works that are included and have credited the sources in my article. I have informed the co-author(s) of the terms of this publishing agreement.
Copyright © Slovenian Society Informatika
